Contents

1   Introduction

This document describes how to perform the health check procedure on the CUDB node.

A health check is performed to verify that no degradations have been introduced into the network after procedures such as reconfiguration, software updates and software upgrades. A health check can also be performed during emergencies to quantify the problems in the network. When changes are made in the network, data used for verification must be collected manually, both before and after the change.

The health check procedures are recommended to be performed before and after a system update/upgrade, a normal backup, or during periodic maintenance. They can also be run as basic normality checks.

1.1   Revision Information

1.2   Prerequisites

This section describes the prerequisites for performing the health check procedure.

1.2.1   Documents

Before starting this procedure, ensure that the following information or documents are available:

• This document.
• All the documents listed in the References section, see References, Reference List.

1.2.2   Tools

The following tool can be used for the health check:

• UDC Cockpit

1.2.3   Conditions

Before health check can be performed, the following conditions must be met.

The users who perform the Health Check must possess the followings:

• Basic knowledge of the CUDB System.
• Knowledge of the IP addresses for the CUDB nodes.
• Required passwords. Refer to CUDB Users and Passwords, Reference [1] CUDB Users and Passwords for more information on the required users and passwords.

1.3   Related Information

Definition and explanation of acronyms and terminology, trademark information, and typographic conventions can be found in the following documents:

• CUDB Glossary of Terms and Acronyms, Reference [2]
• Trademark Information, Reference [3]
• Typographic Conventions, Reference [4]

2   Health Check Tasks

To determine the node health, the following must be verified or checked:

• The CUDB software version.
• The Blackboard Coordination (BC) cluster status.
• The System Monitor (SM) status.
• The status of the master and slave clusters.
• The status of the Data Store (DS).
• The status of replication and the active channel.
• The raised alarms.
• The connection to the database cluster servers.
• The running CUDB processes.

These tasks can be checked by executing the cudbSystemStatus command. See Section 3.1 for more information about the command.

3   Health Check Procedure

This section describes the procedures for determining the health of the node.

3.1   Status of the CUDB System

The cudbSystemStatus command is the primary tool for the Health Check. Use this command to get status information on the CUDB system.

To execute a normality health check on the CUDB system, run the following command:

cudbSystemStatus

For use and help regarding the command, run the following command:

cudbSystemStatus -h

3.1.1   Output of cudbSystemStatus

This section provides an example output of the cudbSystemStatus command. The output is divided into several blocks, each having an explanation below the fault examples. Refer to CUDB Node Commands and Parameters, Reference [5] for more information about the command.

Example 1 shows the first part of the command output, listing the available System Monitor (SM) leaders:

CUDB10 SC_2_1# cudbSystemStatus

Execution date: Thu Aug 30 13:52:48 CEST 2012

CUDB Software Version:
!- CUDB DESIGN DISTRIBUTION: CUDB13B CXP9020214/6 R1B549

Checking BC clusters:

[Site 1]

	SM leader: Node 10 OAM1

	Node 10
		BC server in OAM1 ......... running
		BC server in OAM2 ......... running (Leader)
		BC server in PL2 ......... running

[Site 2]

	SM leader: Node 11 OAM1

	Node 11
		BC server in OAM1 ......... running
		BC server in OAM2 ......... running (Leader)
		BC server in PL2 ......... running

[Site 3]

	SM leader: Node 9 OAM2

	Node 9
		BC server in OAM1 ......... running
		BC server in OAM2 ......... running (Leader)
		BC server in PL2 ......... running

Checking System Monitor BC status in local node:

        SM-BC in OAM1 ......... running
        SM-BC in OAM2 ......... running

Note:  

For CUDB systems deployed on native BSP 8100 with different hardware types, the output will show the used hardware types immediately after the CUDB software version block, as shown in Example 2.

CUDB10 SC_2_1# cudbSystemStatus

Execution date: Sat Mar 12 14:43:28 CET 2016

CUDB Software Version:
!- CUDB DESIGN DISTRIBUTION: CUDB13B CXP9020214/6 R1B549

Checking Hardware Type:
This system is working on following hardware types: EBS_GEP3, EBS_GEP5.
...

Example 3 shows the second part of the command output, listing the status of the Processing Layer Database (PLDB) and Data Store Unit Group (DSG) clusters:

Checking Clusters status:
Node 9:
	PL Cluster (29%) .............................OK
	DSG1 Cluster (23%) ...........................OK
	DSG255 Cluster (23%) .........................OK
Node 10:
	PL Cluster (29%) .............................OK
	DSG1 Cluster (23%) ...........................OK
	DSG255 Cluster (23%) .........................OK
Node 11:
	PL Cluster (29%) .............................OK
	DSG1 Cluster (22%) ...........................OK
	DSG255 Cluster (22%) .........................OK

Checking NDB status:
	PL NDB's (2/2) ...............................OK
	DS1 NDB's (2/2) ..............................OK
	DS2 NDB's (2/2) ..............................OK

If any of the clusters shows a different value than OK, check the active alarms and refer to the related Alarm Operating Instructions (OPIs).

Example 4 shows the third part of the command output, listing the status of the replication channels in the system.

Checking Replication Channels in the System:
	Node    |  9  | 10  | 11  
	==========================
	PLDB ___|__S1_|__S1_|__M__
	DSG 1 __|__S1_|__S1_|__M__
	DSG 255 |__S1_|__S1_|__M__
	
Printing Detailed Replication Status for the Slave Replicas:
Node 9:
	 Replication in DSG0(Chan=1)   .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
	 Replication in DSG1(Chan=1)   .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
	 Replication in DSG255(Chan=1) .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
Node 10:
	 Replication in DSG0(Chan=1)   .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
	 Replication in DSG1(Chan=1)   .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
	 Replication in DSG255(Chan=1) .... Up -- Delay = 0.0 seconds,   no. of pending changes = 0
Node 11:
	 There are no Slave clusters

See Example 9 for an example of replication channel faults.

Example 5 shows the fourth part of the output, listing the active alarms raised by the system:

Printing Alarms...  
[Aug 30 12:50:05]( Preventive Maintenance  Logchecker has \
found major error(s). )

If the Printing Alarms segment shows any alarms, check their related Alarm OPIs.

Example 6 shows the fifth part of the output, listing the status of the database cluster server connections:

Checking MySQL server connection: 
        MySQL Master Servers connection ..............OK
        MySQL Slave Servers connection ...............OK
        MySQL Access Servers connection ..............OK

If any of the database cluster connections shows a value other than OK, check the active alarms and follow the related Alarm OPIs. See Example 10 for an example of database cluster server connection faults.

Example 7 shows the sixth part of the command output, listing the status of the Cluster Supervisor and the Blackboard Coordination (BC) SM:

Checking Process:
OAMs..................

        Cluster Supervisor............................Running 
        System Monitor BC.............................Running 

Example 8 shows the final part of the command output, listing the status of the running processes:

OAMs..................
        Cluster Supervisor............................Running
        System Monitor BC.............................Running
        Reconciliation process........................Running in: OAM2
        Smp-client....................................Running
        Management Server Process (ndb_mgmd)..........Running
        KeepAlive process.............................Running
        ESA...........................................Running
        LDAP counter..................................Running in: OAM1
        Log Handler process...........................Running
        KpiCentral process............................Running in: OAM1
PLs................
        Storage Engine process (ndbd).................Running
        LDAP FE.......................................Running
        KeepAlive process.............................Running
        MySQL server process (Master).................Running
        MySQL server process (Slave)..................Running
        MySQL server process (Access).................Running
        CudbNotifications process.....................Running
        LDAP FE Monitor process.......................Running
DSs............................
        Storage Engine process (ndbd).................Running
        LDAP FE.......................................Running
        KeepAlive process.............................Running
        MySQL server process (Master).................Running
        MySQL server process (Slave)..................Running
        MySQL server process (Access).................Running
        LDAP FE Monitor process.......................Running

See Example 11 for an example of process faults.

3.1.1.1   Fault Examples

This section shows examples of the faults that can be indicated in the output of the cudbSystemStatus command. If any of these faults occurs, contact the next level of support. The examples are as follows:

               |Node 45  |Node 46
==============================
[-E-]PLDB _____|____Xm___|____Xu___
[-E-]DSG 1 ____|____Xm___|____Xu___
[-E-]DSG 2 ____|____Xm___|____Xu___

Printing Detailed Replication Status for the Slave Replicas:
Node 45:
	 There are no Slave clusters
Node 46:
	 There are no Slave clusters

[-W-]   MySQL Slave Server connection Fault in.....: PL_2_3

OAMs..................
[-W-]	Cluster Supervisor............................Not running in: OAM2  
[-W-]	System Monitor BC.............................Not running in: OAM2  
[-W-]	Reconciliation process........................Not running in: OAM1  OAM2 
[-W-]	Smp-client....................................Not running in: OAM2  
[-W-]	Management Server Process (ndb_mgmd)..........Not running in: OAM2  
[-W-]	KeepAlive process.............................Not running in: OAM2  
[-W-]	ESA...........................................Not running in: OAM2  
[-W-]	LDAP counter..................................Not running in: OAM1  OAM2  
[-W-]	Log Handler process...........................Not running in: OAM2  
[-W-]	KpiCentral process............................Not running in: OAM1  
PLs................
[-W-]	Storage Engine process (ndbd).................Not running in: PL0  
[-W-]	LDAP FE.......................................Not running in: PL0  
[-W-]	KeepAlive process.............................Not running in: PL0  
[-W-]	MySQL server process (Master).................Not running in: PL0  
[-W-]	MySQL server process (Slave)..................Not running in: PL0  
[-W-]	MySQL server process (Access).................Not running in: PL0  
[-W-]	CudbNotifications process.....................Not running in: PL0  
[-W-]	LDAP FE Monitor process.......................Not running in: PL0  
DSs............................
[-W-]	Storage Engine process (ndbd).................Not running in: DS2_1  
[-W-]	LDAP FE.......................................Not running in: DS2_1  
[-W-]	KeepAlive process.............................Not running in: DS2_1  
[-W-]	MySQL server process (Master).................Not running in: DS2_1  
[-W-]	MySQL server process (Slave)..................Not running in: DS2_1  
[-W-]	MySQL server process (Access).................Not running in: DS2_1  
[-W-]	LDAP FE Monitor process.......................Not running in: DS2_1  

3.2   UDC Cockpit Tool for Health Check

To follow present and recall earlier system status and performance information of CUDB nodes, use the UDC Cockpit tool. This is a monitoring application, which presents collected data on a single, web-based GUI.

3.3   Detailed Health Check Procedures

This section describes the detailed procedures for determining the health of the node.

3.3.1   Checking Active Alarms

The alarm list can be checked through the Operational Support System (OSS) or by executing the fmactivealarms command as shown in the example below:

SC_2_1# fmactivealarms

3.3.2   Checking ESA Processes

Use the esa status command to check if the ESA processes are running in both SCs. All ESA agents must be running. See the example below on how to run the command and what the expected output looks like:

SC_2_1# esa status

Expected output:

[info] ESA Sub Agent is running.
[info] ESA Master Agent is running.
[info] ESA PM Agent is running.

SC_2_1# ssh OAM2 esa status

Expected output:

[info] ESA Sub Agent is running.
[info] ESA Master Agent is running.
[info] ESA PM Agent is running.

Check the ESA cluster status with the following command on any of the SCs:

esaclusterstatus

The expected output must look similar to the example below. One SC must be in M state, while the other in (M) state:

M * OAM1 10.22.0.1
(M) OAM2 10.22.0.2

The description of the different states are:

3.3.3   Checking Database Cluster Load

The drop ratio of the PLDB and the different DS clusters can be used to estimate the cluster load. Use the pmreadcounter command as follows to see the value of the PLDB drop ratio:

SC_2_1# pmreadcounter | grep DropRatios | grep -i PLDB

The same command is used to check the drop ratio of a specific DS - the only difference is that instead of PLDB, the target DS is defined:

SC_2_1# pmreadcounter | grep DropRatios | grep -i DSX

In the above example, X stands for the target DS number.

In case the drop ratio is greater than or equal to 5 (that is, the amount of traffic drop is 5% or higher), contact the next level of support.

3.3.4   Checking Notifications Traffic

To check if the notifications traffic is being executed, check the following log located in the syslog of the payload blade or VM where the CudbNotifications process is running:

/var/log/PL_2_x/messages

To find the payload blade or VM running the CudbNotifications process, execute the following command:

CUDB3 SC_2_1# cudbHaState | grep NOTIF

The expected output must look similar as follows:

saAmfSISUHAState."safSu=PL-3,safSg=2N,safApp=ERIC-CUDB_SOAP_NOTIFIER"."safSi=2N-1": standby(2)
saAmfSISUHAState."safSu=PL-4,safSg=2N,safApp=ERIC-CUDB_SOAP_NOTIFIER"."safSi=2N-1": active(1)

Refer to CUDB Notifications, Reference [6] for more information.

3.3.5   Checking CPU Load

Check the value of kpiClusterLoad and kpiRatioDroppedCluster counters in the associated 3GPP xml files to determine the CPU load of all database clusters in the CUDB node. For more information about the counters and the files, refer to CUDB Counters List, Reference [7] and CUDB Performance Guide, Reference [8].

3.3.6   Checking Database Consistency

There are two methods to perform a consistency check:

To perform a lightweight consistency check, use the cudbCheckConsistency command to check database consistency between database clusters (that is, between the master PLDB or DSG replicas and their slaves) as follows:

SC_2_1# cudbCheckConsistency

To run an in-depth consistency check on the data of LDAP entry attributes between database clusters (that is, between the master PLDB or DSG replicas and their slaves), use the cudbConsistencyMgr command as follows:

SC_2_1# cudbConsistencyMgr --order ms --node <nodeid> {--dsg <dsgid> | --pl}

To find out how the CUDB Consistency Check function works and how it can be used, refer to the "Consistency Check" section of CUDB Consistency Check, Reference [10].

4   Problem Reporting

For any abnormal situation, refer to CUDB Troubleshooting Guide, Reference [9].

If the problem still exists, report it to the next level of support.

It is also important to collect the related data. For information on how to collect the data, refer to Data Collection Guideline for CUDB, Reference [11].

Glossary

For the terms, definitions, acronyms and abbreviations used in this document, refer to CUDB Glossary of Terms and Acronyms, Reference [2].

Reference List